Skip to main navigation Skip to search Skip to main content

Evaluating the power efficiency of deep learning inference on embedded GPU systems

  • Mahidol University
  • Natl. Inst. Adv. Indust. Sci. T.

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

40 Citations (Scopus)

Abstract

Deep learning inference on embedded systems requires not only high throughput but also low power consumption. To address this challenge, this paper evaluates the power efficiency of image recognition with YOLO, a real-time object detection algorithm, on the latest NVIDIA embedded GPU systems: Jetson TX1 and TX2. For this evaluation, we deployed the Low-Power Image Recognition Challenge (LPIRC) system and integrated YOLO, a power meter, and target hardware into the system. The experimental results show that Jetson TX2 with Max-N mode has the highest throughput; Jetson TX2 with Max-Q mode has the highest power efficiency. These findings indicate it is possible to adjust the trade-off relationship of throughput and power efficiency in Jetson TX2. Therefore, Jetson TX2 has advantages for image recognition on embedded systems more than Jetson TX1 and a PC server with NVIDIA Tesla P40.

Original languageEnglish
Title of host publicationProceeding of 2017 2nd International Conference on Information Technology, INCIT 2017
EditorsWudhichart Sawangphol, Jarernsri L. Mitrpanont
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1-5
Number of pages5
ISBN (Electronic)9781538614310
DOIs
Publication statusPublished - 1 Jul 2017
Event2nd International Conference on Information Technology, INCIT 2017 - Nakhon Pathom, Thailand
Duration: 2 Nov 20173 Nov 2017

Publication series

NameProceeding of 2017 2nd International Conference on Information Technology, INCIT 2017
Volume2018-January

Conference

Conference2nd International Conference on Information Technology, INCIT 2017
Country/TerritoryThailand
CityNakhon Pathom
Period2/11/173/11/17

Keywords

  • deep learning
  • embedded GPU system
  • low-power image recognition
  • object detection
  • power efficiency

Fingerprint

Dive into the research topics of 'Evaluating the power efficiency of deep learning inference on embedded GPU systems'. Together they form a unique fingerprint.

Cite this